Hello Guest it is March 19, 2024, 01:47:30 AM

Author Topic: Physical buttons for plasma  (Read 154324 times)

0 Members and 1 Guest are viewing this topic.

Offline beefy

*
  •  138 138
    • View Profile
Re: Physical buttons for plasma
« Reply #150 on: January 29, 2016, 03:02:35 PM »
Talking about overshoot, that's a problem that will be made a lot worse by having any delay in the control loop.

Has anyone ever came across any tests done to measure how long it takes from Mach3 receiving a THC UP/DOWN input signal to outputting the step/direction command to the stepper drive.

I'd assume that was programmed to be as "instant" as possible.

Keith.

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Physical buttons for plasma
« Reply #151 on: January 29, 2016, 03:03:31 PM »
What am I doing this for?

I suppose the initial idea was just to get the THC as responsive as possible, whether there would be any benefit  over running the THC via Mach to the CSMIO I don't know, maybe it would actually be faster doing away with the Mini THC and using the CSMIO's built in THC function. Just don't know at this stage.


I have emailed Denis to see what the response rate of the MiniTHC is, will see what the reply is, might be too slow to take any advantage that the THC direct to  servo would give (if it even would anyway :D )

Made a board up last night with a couple of optos and just hooked it up to see what it was like, seemed fine and I never even let any smoke out, so that is a bonus :)
Here is a vid of it working, obviously it doesn't show much except the servo getting to speed very quicky in either direction.

https://youtu.be/jWw6fMmjF1Q
Hood

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: Physical buttons for plasma
« Reply #152 on: January 29, 2016, 04:26:18 PM »
Keith

The THC up/down inputs I believe are sampled and processed  at Kernal speed so that part is Quick. AND yes there ARE a few advantages of using a faster kernal speed (;-)

AND I believe that your THC sample rate will determen how fast it can react to a voltage change. Too Slow and it caps teh speed of teh Z  early ??

Hood I could not see much of the motor but that is a quick set of fingers you have there.

(;-) TP
« Last Edit: January 29, 2016, 04:29:58 PM by BR549 »

Offline Hood

*
  •  25,835 25,835
  • Carnoustie, Scotland
    • View Profile
Re: Physical buttons for plasma
« Reply #153 on: January 29, 2016, 05:20:36 PM »
LOL TP, yes afraid it was not easy trying to turn the pot, hold the camera, hold the motor down and get the THC, pot and motor all in the same shot. At least you can see the pulley though :D

Oh and you got a good shot of the tidy section of my desk, and I am not kidding about that being the tidy section BTW :D


Hood

Offline beefy

*
  •  138 138
    • View Profile
Re: Physical buttons for plasma
« Reply #154 on: January 29, 2016, 05:54:53 PM »
Keith

I believe that your THC sample rate will determen how fast it can react to a voltage change. Too Slow and it caps teh speed of teh Z  early ??

(;-) TP

Electronics and computers can be super fast and assuming that is so, then the real limiting factor is mechanical, so let's have a think about this:

My theory is that the maximum USEABLE rate of a THC is completely determined by how fast a Z-axis can change direction then change back again. If you had a THC sample rate of say 50 times/sec then the worst case scenario is for 1/50 second the Z is being told to go up then for the next 1/50 second the Z is being told to go down. How many Z axis are capable of taking advantage of such demands.

I can easily bump up my sample rate to 100/sec and that's still giving me 10 milliseconds per sample to do all the calculations, compare with SET VOLTS, etc and give a THC UP/DOWN signal, certainly doable with my microcontroller.

BUT is there any benefit to reacting in 1/100th second when I can't see a Z axis being mechanically capable of taking advantage of that speed. I heard that a humming birds wings flap at about 50 times a second and when I see those little guys I can't imagine a Z axis competing with them.

And how long would be considered too long to wait for a THC UP/DOWN signal when the volts had gone outside of the SET VOLTS deadband range. My reasoning for NOT wanting to sample at such a high rate is that I can sample the voltage for longer periods and get better averaging of the noisy spiky plasma voltage.

Do you see my reasoning ?

Keith.
« Last Edit: January 29, 2016, 06:02:47 PM by beefy »

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: Physical buttons for plasma
« Reply #155 on: January 29, 2016, 06:36:28 PM »
YES BUT if you sample to low it effects teh quality of teh calculated values. I am SURE you are aware of the Niquest theorum and sampling rates.

The reaction time needed for THC is directly related to teh Cutting speed AND rate of change in the Material surface heights.  Cutting thin corragated sheetsgoods at high speeds can be a challange to get right.  So far as Ian once mentioned most of the cuts you see demonstarting it done are still fairly SLOW cuts. I had to go back and watch a LOT of the videos. IF you look at teh sheet it is a fairly thick gauge of metal (;-)

Yes Ian I do read your post(;-)

Just a thought, (;-) TP
« Last Edit: January 29, 2016, 06:45:44 PM by BR549 »

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Physical buttons for plasma
« Reply #156 on: January 30, 2016, 04:44:19 AM »
The videos on the 'tube seem to very low speed, impressive but low speed for the thickness being cut.

I was also pondering what happens to the voltage when the tip of the torch shied is about 8mm round on my PM45 and would be touching the slope of the metal - the arc will be longer as the torch can't physically get close enough, do you have to fudge the voltage and run with arc-stretch on heavily sloped metal.

Tilt the torch on the fly so its perpendicular all the time then on the up/down slopes you can vary the X/Y velocity to adjust height :)

The ultimate aim is to maintain arc length so is there a vision system that can do that? Forget trying to sample arc voltage and use a system that can look ahead and plan the move???

Too advanced??

;)

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Physical buttons for plasma
« Reply #157 on: January 30, 2016, 04:18:21 PM »
HI Dave.

IF the arc volts go uP then the torch goes down

If the Arc volts go down the torch goes up

Anti corner lock blocks torch down ( Mach3 function)

Anti Gap lock blocks torch down across a void ( THC control function)

What else would you think it needs ?

I would start with the Corner lock at %90 it will vary as to your machine and it motion charectoristics

(;-) TP

Just a quickie...
I was fine tuning some changes to my system today and i happened to notice that Mach anti-dive wasn't even turned on, never noticed the little button and flashing yellow led before - probably been off from day one!

D'oh!

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: Physical buttons for plasma
« Reply #158 on: January 30, 2016, 04:43:07 PM »
hIYA Dave you MAY want to readup about Tip Saver IF I remember back that far correctly I think it did both functions AND Tom recommended NOT to run both at teh same time.

(;-) TP

Offline Davek0974

*
  •  2,606 2,606
    • View Profile
Re: Physical buttons for plasma
« Reply #159 on: January 30, 2016, 04:51:37 PM »
hmm i'll double check that, i have all the numerous candcnc manuals ;)

it read like it was ok as mach anti dive was speed related and candcnc tip saver was voltage-spike related, does the voltage rise as speed drops?

if so then i guess that may be why it was turned off, don't recall reading any warnings about it though.